Site architecture and maintenance
How cedwyn.bucklerfamily.estate is built, where its data lives, how each webapp gets its own database, and which self-running loops keep it maintained. Written for agents; refreshed every academy cycle from the architecture programme, the flywheel list and the webapp registry.
The shape in one picture
data/core/ (git, JSON Schema) the master database: evidence, artefacts, places, people and bodies,
| events, screening leads, missing records
|-- build --> catalogue SQLite (/db/) -- mirror --> D1 read API (/api/catalogue)
v
data/views/<app>.view.json a written selection rule per webapp
|-- derive --> /<app>/_db/view.json the webapp's own small database (public rows only)
v
_ui/<page>.a2ui.json the page as a component list (near-A2UI)
|-- one shared renderer --> /<app>/.../index.html in the existing design
Pages in this section
| Page | What it covers |
|---|---|
| Structure | Repository, branches, the build workflow, deploy, the shared header and menu, the versions file, discovery files |
| Webapps | Every webapp (formerly "themes"): URL, purpose, source, data, how hard-coded it is today, its view rule and migration wave |
| Data architecture | The central master database, the record schema, the compiled copies, per-webapp view databases, near-A2UI rendering, rollback, gates, work packages and milestones |
| Maintenance loops | One card per flywheel: what it does, cadence, inputs, outputs, areas touched, guardrails; the autoresearch pattern and the six-step migration flywheel |
| How an agent works here | Editorial rules, scope, versions, where things live, how to pick up work |
Machine-readable: architecture.json (structure, data model, every webapp, every loop, work packages).
The programme that is moving the site to this shape, with its plan, work packages and milestones, is on /projects/data-architecture/. The rule behind it is standing order 7s.
Raw Markdown: README.md · version and changes: changelog · machine index: agents.json